TOHOKU UNIVERSITY Researchers

Details of the Researcher

Home

日本語 English

Tomo Miyazaki

Section

Graduate School of Engineering

Job title

Associate Professor

Degree

博士（工学）（東北大学）
修士（工学）（東北大学）

researchmap

https://researchmap.jp/tomo_miyazaki

J-GLOBAL ID

201601004556825881

ORCID

https://orcid.org/0000-0001-5205-0542

Research History 4

2024/05 - Present

Tohoku University　Associate Professor
2015/04 - 2024/04

東北大学　Assitant Professor
2013/01 - 2015/03

東北大学　Researcher
2011/04 - 2012/12

（株）日立製作所ディフェンスシステム社

Education 2

Tohoku University　Graduate School, Division of Engineering　Electrical and Communication Engineering

- 2011/03/25
Yamagata University　Faculty of Engineering　情報科学科

- 2006/03/25

Committee Memberships 2

画像工学研究会　専門委員

2019/06 - 2023/06
パターン認識・メディア理解研究会　専門委員

2015/06 - 2021/06

Professional Memberships 2

IEEE
電子情報通信学会

Research Interests 3

Document Analysis
image processing
Pattern recognition

Research Areas 1

Informatics / Perceptual information processing /

Awards 6

インタラクティブ発表賞

2024/08　Domain Generalizable Multi-Targeted Adversarial Attack using Dynamic Loss Weighting
インタラクティブ発表賞

2024/08　第27回画像の認識・理解シンポジウム　生成画像を用いた工業画像マルチクラス分類と異常検知アルゴリズム性能との比較検証
石田實記念財団研究奨励賞

2021/11　一般財団法人石田實記念財団
The Eighth International Conferences on Pervasive Patterns and Applications, Best Paper Award

2016/03/22　The Eighth International Conferences on Pervasive Patterns and Applications
The Eighth International Conferences on Creative Content Technologies, Best Paper Award

2016/03/22　The Eighth International Conference on Creative Content Technologies
IEEE Sendai Section, The Best Paper Prize

2007/12/07　IEEE Sendai Section

Show all ︎Show 5

Papers 86

Semantically-Guided Image Compression for Enhanced Perceptual Quality at Extremely Low Bitrates Peer-reviewed

Shoma Iwai, Tomo Miyazaki, Shinichiro Omachi

IEEE Access　12　100057-100072　2024
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
DOI： 10.1109/access.2024.3430322 　

eISSN： 2169-3536
Japanese historical character recognition by focusing on character parts

Takuru Ishikawa, Tomo Miyazaki, Shinichiro Omachi

Pattern Recognition　110181-110181　2023/12
Publisher: Elsevier BV
DOI： 10.1016/j.patcog.2023.110181 　

ISSN： 0031-3203
A Scene-Text Synthesis Engine Achieved Through Learning from Decomposed Real-World Data International-journal Peer-reviewed

Zhengmi Tang, Tomo Miyazaki, Shinichiro Omachi

IEEE Transactions on Image Processing　2023/11

DOI： 10.1109/TIP.2023.3326685 　
Deep Image Compression Using Scene Text Quality Assessment Peer-reviewed

Shohei Uchigasaki, Tomo Miyazaki, Shinichiro Omachi

Pattern Recognition　109696-109696　2023/05
Publisher: Elsevier BV
DOI： 10.1016/j.patcog.2023.109696 　

ISSN： 0031-3203
Important Region Estimation Using Image Captioning Peer-reviewed

Taku Suzuki, Daisuke Sato, Yoshihiro Sugaya, Tomo Miyazaki, Shinichiro Omachi

IEEE Access　10　105546-105555　2022/10
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
DOI： 10.1109/access.2022.3211260 　

eISSN： 2169-3536
Stroke-Based Scene Text Erasing Using Synthetic Data for Training International-journal Peer-reviewed

Zhengmi Tang, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

IEEE Transactions on Image Processing　30　9306-9320　2021/11
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
DOI： 10.1109/tip.2021.3125260 　

ISSN： 1057-7149

eISSN： 1941-0042
A Framework for Estimating Gaze Point Information for Location-Based Services Peer-reviewed

Junpei Masuho, Tomo Miyazaki, Yoshihiro Sugaya, Masako Omachi, Shinichiro Omachi

IEEE Transactions on Vehicular Technology　70　(9)　8468-8477　2021/09
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
DOI： 10.1109/tvt.2021.3101932 　

ISSN： 0018-9545

eISSN： 1939-9359
Graph Neural Networks with Multiple Feature Extraction Paths for Chemical Property Estimation International-journal Peer-reviewed

Sho Ishida, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

Molecules　26　(11)　3125-3125　2021/05/24
Publisher: MDPI AG
DOI： 10.3390/molecules26113125 　

eISSN： 1420-3049

More details Close

Feature extraction is essential for chemical property estimation of molecules using machine learning. Recently, graph neural networks have attracted attention for feature extraction from molecules. However, existing methods focus only on specific structural information, such as node relationship. In this paper, we propose a novel graph convolutional neural network that performs feature extraction with simultaneously considering multiple structures. Specifically, we propose feature extraction paths specialized in node, edge, and three-dimensional structures. Moreover, we propose an attention mechanism to aggregate the features extracted by the paths. The attention aggregation enables us to select useful features dynamically. The experimental results showed that the proposed method outperformed previous methods.
Automatic Generation of Typographic Font From Small Font Subset International-journal Peer-reviewed

Tomo Miyazaki, Tatsunori Tsuchiya, Yoshihiro Sugaya, Shinichiro Omachi, Masakazu Iwamura, Seiichi Uchida, Koichi Kise

IEEE Computer Graphics and Applications　40　(1)　99-111　2020/01/01
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
DOI： 10.1109/mcg.2019.2931431 　

ISSN： 0272-1716

eISSN： 1558-1756
Object-Based Video Coding by Visual Saliency and Temporal Correlation International-journal Peer-reviewed

Kazuya Ogasawara, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

IEEE Transactions on Emerging Topics in Computing　8　(1)　168-178　2020/01/01
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
DOI： 10.1109/tetc.2017.2695640 　

eISSN： 2376-4562
Automatic Mackerel Sorting Machine using Global and Local Features International-journal Peer-reviewed

Yoshito Nagaoka, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

IEEE Access　7　(1)　63767-63777　2019/12

DOI： 10.1109/ACCESS.2019.2917554 　
Structural Data Recognition With Graph Model Boosting International-journal Peer-reviewed

Tomo Miyazaki, Shinichiro Omachi

IEEE Access　6　63606-63618　2018
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
DOI： 10.1109/access.2018.2876860 　

eISSN： 2169-3536
VQ-STE: Scene text erasing with mask refinement and vector-quantized texture dictionary

Zhengmi Tang, Tomo Miyazaki, Zhijie Wang, Yongsong Huang, Jonathan Pradana Mailoa, Shinichiro Omachi

Knowledge-Based Systems　315　113306-113306　2025/04
Publisher: Elsevier BV
DOI： 10.1016/j.knosys.2025.113306 　

ISSN： 0950-7051
Texture and noise dual adaptation for infrared image super-resolution

Yongsong Huang, Tomo Miyazaki, Xiaofeng Liu, Yafei Dong, Shinichiro Omachi

Pattern Recognition　111449-111449　2025/02
Publisher: Elsevier BV
DOI： 10.1016/j.patcog.2025.111449 　

ISSN： 0031-3203
Dual-Conditioned Training to Exploit Pre-Trained Codebook-Based Generative Model in Image Compression International-journal Peer-reviewed

Shoma Iwai, Tomo Miyazaki, Shinichiro Omachi

IEEE Access　12　198184-198200　2024/12

DOI： 10.1109/ACCESS.2024.3522238 　
Lightweight Reference-Based Video Super-Resolution Using Deformable Convolution

Tomo Miyazaki, Zirui Guo, Shinichiro Omachi

Information　15　(11)　718-718　2024/11/08
Publisher: MDPI AG
DOI： 10.3390/info15110718 　

eISSN： 2078-2489

More details Close

Super-resolution is a technique for generating a high-resolution image or video from a low-resolution counterpart by predicting natural and realistic texture information. It has various applications such as medical image analysis, surveillance, remote sensing, etc. However, traditional single-image super-resolution methods can lead to a blurry visual effect. Reference-based super-resolution methods have been proposed to recover detailed information accurately. In reference-based methods, a high-resolution image is also used as a reference in addition to the low-resolution input image. Reference-based methods aim at transferring high-resolution textures from the reference image to produce visually pleasing results. However, it requires texture alignment between low-resolution and reference images, which generally requires a lot of time and memory. This paper proposes a lightweight reference-based video super-resolution method using deformable convolution. The proposed method makes the reference-based super-resolution a technology that can be easily used even in environments with limited computational resources. To verify the effectiveness of the proposed method, we conducted experiments to compare the proposed method with baseline methods in two aspects: runtime and memory usage, in addition to accuracy. The experimental results showed that the proposed method restored a high-quality super-resolved image from a very low-resolution level in 0.0138 s using two NVIDIA RTX 2080 GPUs, much faster than the representative method.
TAMC: Textual Alignment and Masked Consistency for Open-Vocabulary 3D Scene Understanding Peer-reviewed

Juan Wang, Zhijie Wang, Tomo Miyazaki, Yaohou Fan, Shinichiro Omachi

Sensors　24　(19)　6166-6166　2024/09/24
Publisher: MDPI AG
DOI： 10.3390/s24196166 　

eISSN： 1424-8220

More details Close

Three-dimensional (3D) Scene Understanding achieves environmental perception by extracting and analyzing point cloud data with wide applications including virtual reality, robotics, etc. Previous methods align the 2D image feature from a pre-trained CLIP model and the 3D point cloud feature for the open vocabulary scene understanding ability. We believe that existing methods have the following two deficiencies: (1) the 3D feature extraction process ignores the challenges of real scenarios, i.e., point cloud data are very sparse and even incomplete; (2) the training stage lacks direct text supervision, leading to inconsistency with the inference stage. To address the first issue, we employ a Masked Consistency training policy. Specifically, during the alignment of 3D and 2D features, we mask some 3D features to force the model to understand the entire scene using only partial 3D features. For the second issue, we generate pseudo-text labels and align them with the 3D features during the training process. In particular, we first generate a description for each 2D image belonging to the same 3D scene and then use a summarization model to fuse these descriptions into a single description of the scene. Subsequently, we align 2D-3D features and 3D-text features simultaneously during training. Massive experiments demonstrate the effectiveness of our method, outperforming state-of-the-art approaches.
Machine learning-based identification of the risk factors for postoperative nausea and vomiting in adults Peer-reviewed

Hiroshi Hoshijima, Tomo Miyazaki, Yuto Mitsui, Shinichiro Omachi, Masanori Yamauchi, Kentaro Mizuta

PLOS ONE　19　(8)　1-15　2024/08/15
Publisher: Public Library of Science (PLoS)
DOI： 10.1371/journal.pone.0308755 　

eISSN： 1932-6203

More details Close

Postoperative nausea and vomiting (PONV) is a common adverse effect of anesthesia. Identifying risk factors for PONV is crucial because it is associated with a longer stay in the post-anesthesia care unit, readmissions, and perioperative costs. This retrospective study used artificial intelligence to analyze data of 37,548 adult patients (aged ≥20 years) who underwent surgery under general anesthesia at Tohoku University Hospital from January 1, 2010 to December 31, 2019. To evaluate PONV, patients who experienced nausea and/or vomiting or used antiemetics within 24 hours after surgery were extracted from postoperative medical and nursing records. We create a model that predicts probability of PONV using the gradient tree boosting model, which is a widely used machine learning algorithm in many applications due to its efficiency and accuracy. The model implementation used the LightGBM framework. Data were available for 33,676 patients. Total blood loss was identified as the strongest contributor to PONV, followed by sex, total infusion volume, and patient’s age. Other identified risk factors were duration of surgery (60–400 min), no blood transfusion, use of desflurane for maintenance of anesthesia, laparoscopic surgery, lateral positioning during surgery, propofol not used for maintenance of anesthesia, and epidural anesthesia at the lumbar level. The duration of anesthesia and the use of either sevoflurane or fentanyl were not identified as risk factors for PONV. We used artificial intelligence to evaluate the extent to which risk factors for PONV contribute to the development of PONV. Intraoperative total blood loss was identified as the potential risk factor most strongly associated with PONV, although it may correlate with duration of surgery, and insufficient circulating blood volume. The use of sevoflurane and fentanyl and the anesthesia time were not identified as risk factors for PONV in this study.
Learn from orientation prior for radiograph super-resolution: Orientation operator transformer International-journal International-coauthorship Peer-reviewed

Yongsong Huang, Tomo Miyazaki, Xiaofeng Liu, Kaiyuan Jiang, Zhengmi Tang, Shinichiro Omachi

Computer Methods and Programs in Biomedicine　245　2024/03

DOI： 10.1016/j.cmpb.2023.108000 　
IRSRMamba: Infrared Image Super-Resolution via Mamba-based Wavelet Transform Feature Modulation Model.

Yongsong Huang, Tomo Miyazaki, Xiaofeng Liu 0001, Shinichiro Omachi

CoRR　abs/2405.09873　2024

DOI： 10.48550/arXiv.2405.09873 　
JPEG Image Enhancement with Pre-Processing of Color Reduction and Smoothing International-journal Peer-reviewed

Akane Shoda, Tomo Miyazaki, Shinichiro Omachi

Sensors　23　(21)　2023/10

DOI： 10.3390/s23218861 　
Collaborative Indoor Positioning by Localization Comparison at an Encounter Position International-journal Peer-reviewed

Kohei Kageyama, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

Applied Sciences　13　(12)　2023/06

DOI： 10.3390/app13126962 　
Target-oriented Domain Adaptation for Infrared Image Super-Resolution.

Yongsong Huang, Tomo Miyazaki, Xiaofeng Liu, Yafei Dong, Shinichiro Omachi

CoRR　abs/2311.08816　2023

DOI： 10.48550/arXiv.2311.08816 　
Comparative Pulse Shape Discrimination Study for Ca(Br, I)2 Scintillators Using Machine Learning and Conventional Methods Peer-reviewed

M. Yoshino, T. Iida, K. Mizukoshi, T. Miyazaki, K. Kamada, K.J. Kim, A. Yoshikawa

Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment　1045　167626-167626　2023/01
Publisher: Elsevier BV
DOI： 10.1016/j.nima.2022.167626 　

ISSN： 0168-9002
GAN-based Privacy-Conscious Data Augmentation with Finger-Vein Images Peer-reviewed

Yusuke Matsuda, Tomo Miyazaki, Shinichiro Omachi

2022 IEEE International Conference on Artificial Intelligence in Engineering and Technology (IICAIET)　2022/09/13
Publisher: IEEE
DOI： 10.1109/iicaiet55139.2022.9936860 　
Infrared Image Super-Resolution: Systematic Review, and Future Trends.

Yongsong Huang, Tomo Miyazaki, Xiaofeng Liu, Shinichiro Omachi

CoRR　abs/2212.12322　2022

DOI： 10.48550/arXiv.2212.12322 　
Importance Estimation for Scene Texts Using Visual Features Peer-reviewed

Kota OODAIRA, Tomo MIYAZAKI, Yoshihiro SUGAYA, Shinichiro OMACHI

Interdisciplinary Information Sciences　28　(1)　15-23　2022
Publisher: Graduate School of Information Sciences, Tohoku University
DOI： 10.4036/iis.2022.a.06 　

ISSN： 1340-9050

eISSN： 1347-6157
Self Texture Transfer Networks for Low Bitrate Image Compression Peer-reviewed

Shoma Iwai, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops　2021/06

DOI： 10.1109/CVPRW53098.2021.00214 　
Multiple Visual-Semantic Embedding for Video Retrieval from Query Sentence International-journal Peer-reviewed

Huy Manh Nguyen, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

Applied Sciences　11　(7)　3214-3214　2021/04/03
Publisher: MDPI AG
DOI： 10.3390/app11073214 　

eISSN： 2076-3417

More details Close

Visual-semantic embedding aims to learn a joint embedding space where related video and sentence instances are located close to each other. Most existing methods put instances in a single embedding space. However, they struggle to embed instances due to the difficulty of matching visual dynamics in videos to textual features in sentences. A single space is not enough to accommodate various videos and sentences. In this paper, we propose a novel framework that maps instances into multiple individual embedding spaces so that we can capture multiple relationships between instances, leading to compelling video retrieval. We propose to produce a final similarity between instances by fusing similarities measured in each embedding space using a weighted sum strategy. We determine the weights according to a sentence. Therefore, we can flexibly emphasize an embedding space. We conducted sentence-to-video retrieval experiments on a benchmark dataset. The proposed method achieved superior performance, and the results are competitive to state-of-the-art methods. These experimental results demonstrated the effectiveness of the proposed multiple embedding approach compared to existing methods.
Text Detection Using Multi-Stage Region Proposal Network Sensitive to Text Scale International-journal Peer-reviewed

Yoshito Nagaoka, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

Sensors　21　(4)　1232-1232　2021/02/09
Publisher: MDPI AG
DOI： 10.3390/s21041232 　

eISSN： 1424-8220

More details Close

Recently, attention has surged concerning intelligent sensors using text detection. However, there are challenges in detecting small texts. To solve this problem, we propose a novel text detection CNN (convolutional neural network) architecture sensitive to text scale. We extract multi-resolution feature maps in multi-stage convolution layers that have been employed to prevent losing information and maintain the feature size. In addition, we developed the CNN considering the receptive field size to generate proposal stages. The experimental results show the importance of the receptive field size.
Fidelity-Controllable Extreme Image Compression with Generative Adversarial Networks Peer-reviewed

Shoma Iwai, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

Proceedings of the 25th International Conference on Pattern Recognition (ICPR 2020)　8235-8242　2021/01

DOI： 10.1109/ICPR48806.2021.9412185 　
Mackerel Fat Content Estimation Using RGB and Depth Images

Shuya Sano, Tomo Miyazaki, Yoshihiro Sugaya, Naohiro Sekiguchi, Shinichiro Omachi

IEEE Access　9　164060-164069　2021
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
DOI： 10.1109/access.2021.3134260 　

eISSN： 2169-3536
Optical Flow-Based Fast Motion Parameters Estimation for Affine Motion Compensation International-journal Peer-reviewed

Antoine Chauvet, Yoshihiro Sugaya, Tomo Miyazaki, Shinichiro Omachi

Appllied Sciences　10　(2)　2020/01

DOI： 10.3390/app10020729 　
Super Resolution for Multi Frames with 3D Feature Extraction and RNN Prediction Peer-reviewed

Xi Huang, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

2019 International Symposium on Signal Processing Systems Proceedings (SSPS 2019)　2019/09

DOI： 10.1145/3364908.3364909 　
Fast Image Quality Enhancement for HEVC by Postfiltering via Shallow Neural Networks Peer-reviewed

Antoine Chauvet, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

IIEEJ Transactions on Image Electronics and Visual Computing　7　(1)　2-12　2019/06
A Development of Wearable System for Information Providing Applications Using Scene Text Recognition Peer-reviewed

Yoshihiro Sugaya, Kiyoshiro Sakai, Tomo Miyazaki, Shinichiro Omachi

The Journal of the Institute of Image Electronics Engineers of Japan　48　(2)　248-257　2019/04
Text Detection by Faster R-CNN with Multiple Region Proposal Networks Peer-reviewed

Yoshito Nagaoka, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

Proceedings of the International Conference on Document Analysis and Recognition, ICDAR　6　15-20　2018/01/25
Publisher: IEEE Computer Society
DOI： 10.1109/ICDAR.2017.343 　

ISSN： 1520-5363
Glyph-Based Data Augmentation for Accurate Kanji Character Recognition Peer-reviewed

Kenichiro Ofusa, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

Proceedings of the International Conference on Document Analysis and Recognition, ICDAR　1　597-602　2018/01/25
Publisher: IEEE Computer Society
DOI： 10.1109/ICDAR.2017.103 　

ISSN： 1520-5363
Mackerel Classification using Global and Local Features Peer-reviewed

Yoshito Nagaoka, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

Proceedings of the 2018 IEEE 23rd International Conference on Emerging Technologies and Factory Automation　1209-1212　2018
Activity Recognition Using Gazed Text and Viewpoint Information for User Support Systems Peer-reviewed

Shun Chiba, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

Journal of Sensor and Actuator Networks　7　(3)　2018

DOI： 10.3390/jsan7030031 　
Automatic Discrimination between Scomber japonicus and Scomber australasicus by Geometric and Texture Features Peer-reviewed

Airi Kitasato, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

Fishes　3　(3)　2018

DOI： 10.3390/fishes3030026 　
Text retrieval for Japanese historical documents by image generation Peer-reviewed

Chisato Sugawara, Yoshihiro Sugaya, Tomo Miyazaki, Shinichiro Omachi

ACM International Conference Proceeding Series　19-24　2017/11/10
Publisher: Association for Computing Machinery
DOI： 10.1145/3151509.3151512 　
A Preliminary Study on Location Estimation without Preparation using Ceiling Signboard Peer-reviewed

Yoshihiro Sugaya, Kento Takeda, Tomo Miyazaki, Shinichiro Omachi

Proceedings of International Conference on Indoor Positioning and Indoor Navigation　2017/09
Adaptive Post Filter for Reducing Block Artifacts in High Efficiency Video Coding Peer-reviewed

Antoine Chauvet, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

Proceedings - 2016 International Conference on Multimedia Systems and Signal Processing, ICMSSP 2016　22-25　2017/06/09
Publisher: Institute of Electrical and Electronics Engineers Inc.
DOI： 10.1109/ICMSSP.2016.014 　
Development of Wearable System for Translation of Japanese Texts in the Environment Peer-reviewed

Kiyoshiro Sakai, Yoshihiro Sugaya, Tomo Miyazaki, Shinichiro Omachi

Proceedings of International Workshop on Frontiers of Computer Vison　2017/02
Analysis of Floor Map Image in Information Board for Indoor Navigation Peer-reviewed

Tomoya Honto, Yoshihiro Sugaya, Tomo Miyazaki, Shinichiro Omachi

2017 INTERNATIONAL CONFERENCE ON INDOOR POSITIONING AND INDOOR NAVIGATION (IPIN)　2017

DOI： 10.1109/IPIN.2017.8115896 　

ISSN： 2162-7347
Efficient coding for video including text using image generation Peer-reviewed

Yosuke Nozue, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

Journal of Information Processing　24　(2)　330-338　2016/03/15
Publisher: Information Processing Society of Japan
DOI： 10.2197/ipsjjip.24.330 　

ISSN： 1882-6652 0387-5806
Indoor Localization by Map Matching Using One Image of Guide Plate Peer-reviewed

Kento Tonosaki, Yoshihiro Sugaya, Tomo Miyazaki, Shinichiro Omachi

The Eighth International Conferences on Pervasive Patterns and Applications (PATTERNS 2016)　2016/03
Indoor Localization by Map Matching Using One Image of Information Board Peer-reviewed

Kento Tonosaki, Toshihiro Sugaya, Tomo Miyazaki, Shinichiro Omachi

The Eighth International Conferences on Pervasive Patterns and Applications　2016/03
Object-based Video Coding for Arbitrary Shape by Visual Saliency and Temporal Correlation Peer-reviewed

Kazuya Ogasawara, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

The Eighth International Conference on Creative Content Technologies　2016/03
Discrimination of Scomber Japonicus and Scomber Australasicus by Dorsal Fin Length and Fork Length Peer-reviewed

Airi Kitasato, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

The Korea-Japan joint workshop on Frontiers of Computer Vision 2016　338-341　2016/02
Precise Character Region Extraction from Scene Images Using Auxiliary Lines Peer-reviewed

Shiori KAWAMURA, Tomo MIYAZAKI, Yoshihiro SUGAYA, Shinichiro OMACHI

The Journal of the Institute of Image Electronics Engineers of Japan　45　(1)　62-70　2016/02
Publisher: The Institute of Image Electronics Engineers of Japan
DOI： 10.11371/iieej.45.62 　

ISSN： 2186-716X

More details Close

The demand for character recognition in scene images has been growing. However, complex background, non-uniform illumination conditions and shadows are main obstacles for accurate detection and recognition of characters in scene images. For accurate detection, some methods that utilize a supplementary information given by a user with a touch operation using a device such as smart phones or tablets have been proposed. Based on this idea, in this paper, we propose a method for detecting character regions in scene image using auxiliary information given by a user. In the proposed method, the user will designate rough positions of characters in the image by an auxiliary lines and character regions are extracted based on the information. The effectiveness of the proposed method is experimentally evaluated with a public database. It achieves better result for character region extraction if the characters are with complex background and uneven brightness.
Precise character region extraction from scene images using auxiliary lines

Shiori Kawamura, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

Journal of the Institute of Image Electronics Engineers of Japan　45　(1)　62-70　2016
Publisher: Institute of Image Electronics Engineers of Japan
ISSN： 1348-0316 0285-9831
The Journal of the Institute of Image Electronics Engineers of Japan Peer-reviewed

Siori KAWAMURA, Tomo MIYAZAKI, Yoshihiro SUGAYA, Shinichiro OMACHI

画像電子学会誌　45　(1)　62-70　2016/01
スパースコーディングを用いたテキストを含む画像符号化に関する検討

井上慶祐, 宮崎智, 菅谷至寛, 大町真一郎

電子情報通信学会技術研究報告　116　(119)　5-10　2016
Improvement of Map Matching for Indoor Navigation Exploiting Photo of Information Board Peer-reviewed

Kento Tonosaki, Yoshihiro Sugaya, Tomo Miyazaki, Shinichiro Omachi

Proceedings of International Conference on Indoor Positioning and Indoor Navigation　22-26　2016
Graph Model Boosting for Structural Data Recognition Peer-reviewed

Tomo Miyazaki, Shinichiro Omachi

2016 23RD INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION (ICPR)　1707-1712　2016

DOI： 10.1109/ICPR.2016.7899882 　

ISSN： 1051-4651
顕著性マップとGrabCutによる注目物体抽出を用いた動画像符号化

小笠原和也, 宮崎智, 菅谷至寛, 大町真一郎

Proceedings of the 2015 Tohoku-Section Joint Convention of Institutes of Electrical and Information Engineers, Japan　2015　115-115　2015/08/27
Publisher:
DOI： 10.11528/tsjc.2015.0_115 　

More details Close

2D10
電子基板の欠陥検査のための文字認識

吉田大樹, 宮崎智, 菅谷至寛, 大町真一郎

Proceedings of the 2015 Tohoku-Section Joint Convention of Institutes of Electrical and Information Engineers, Japan　2015　128-128　2015/08/27
Publisher:
DOI： 10.11528/tsjc.2015.0_128 　

More details Close

1E06
Accuracy Improvement of Character Recognition Using Generated Samples by Morphing

Shuto Shinbo, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

Proceedings of the 2015 Tohoku-Section Joint Convention of Institutes of Electrical and Information Engineers, Japan　2015　31-31　2015/08/27
Publisher: Organizing Committee of Tohoku-Section Joint Convention of Institutes of Electrical and Information Engineers, Japan
DOI： 10.11528/tsjc.2015.0_31 　

More details Close

2A12
Gaze Detection in Omnidirectional Scene by Iterative Image Matching

Shun Chiba, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

Proceedings of the 2015 Tohoku-Section Joint Convention of Institutes of Electrical and Information Engineers, Japan　2015　30-30　2015/08/27
Publisher: Organizing Committee of Tohoku-Section Joint Convention of Institutes of Electrical and Information Engineers, Japan
DOI： 10.11528/tsjc.2015.0_30 　

More details Close

2A11
Survey of Guide Plates and Fundamental Study of Map Image Analysis for Indoor Navigation

Kento Tonosaki, Yoshihiro Sugaya, Tomo Miyazaki, Shinichiro Omachi

Proceedings of the 2015 Tohoku-Section Joint Convention of Institutes of Electrical and Information Engineers, Japan　2015　127-127　2015/08/27
Publisher: Organizing Committee of Tohoku-Section Joint Convention of Institutes of Electrical and Information Engineers, Japan
DOI： 10.11528/tsjc.2015.0_127 　

More details Close

1E05
屋内ナビゲーションのための案内板画像解析に関する研究

Kento TONOSAKI, Yoshihiro SUGAYA, Tomo MIYAZAKI, Shinichiro OMACHI

電子情報通信学会2015年総合大会学生ポスターセッション　115-115　2015/03
Estimate of Viewpoint Positions in Environment Using Eye Tracker and Omnidirectional Camera

Shun Chiba, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

IEICE Technical Report　114　(454)　101-102　2015/02/12
Publisher: The Institute of Electronics, Information and Communication Engineers

More details Close

PRMU2014-134
Character Detection with Stroke Width Features from Scene Image

Yasushi Oshima, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

IEICE Technical Report　114　(454)　99-100　2015/02/12
Publisher: The Institute of Electronics, Information and Communication Engineers

More details Close

PRMU2014-133
Efflciency Sprit Compression Using lmage Abstraction

Ryosuke Ishimori, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

IEICE Technical Report　114　(454)　97-98　2015/02/12
Publisher: The Institute of Electronics, Information and Communication Engineers

More details Close

PRMU2014-132
Ultra-low Resolution Character Recognition by Utilizing Multiple Frames

Shuhei Toba, Hirotaka Kudo, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

IEICE Technical Report　114　(454)　95-96　2015/02/12
Publisher: The Institute of Electronics, Information and Communication Engineers

More details Close

PRMU2014-131
Ultra-low Resolution Character Recognition System with Pruning Mutual Subspace Method Peer-reviewed

Shuhei Toba, Hirotaka Kudo, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

2015 IEEE INTERNATIONAL CONFERENCE ON CONSUMER ELECTRONICS - TAIWAN (ICCE-TW)　284-285　2015

DOI： 10.1109/ICCE-TW.2015.7216900 　
Estimation of Gazing Points in Environment Using Eye Tracker and Omnidirectional Camera Peer-reviewed

Shun Chiba, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

2015 IEEE INTERNATIONAL CONFERENCE ON CONSUMER ELECTRONICS - TAIWAN (ICCE-TW)　47-48　2015

DOI： 10.1109/ICCE-TW.2015.7217003 　
Finding Stroke Parts for Rough Text Detection in Scene Images with Random Forest Peer-reviewed

Tomo Miyazaki, Shinichiro Omachi

Proceedings of Joint Conference of IWAIT and IFMIA　2015
Split Compression Using Image Abstraction for Object-Based Encoding

Ryosuke Ishimori, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

PCSJ2014　2014/11

More details Close

P-4-11
A Video Coding Method for Text using Character Parameters

Yosuke Nozue, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

PCSJ2014　2014/11

More details Close

P-2-11
A Video Coding Method for Scene Text

Yosuke Nozue, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

Proceedings of the 2014 Tohoku-Section Joint Convention of Institutes of Electrical and Information Engineers, Japan　2014/08/21

More details Close

2A19
Ultra-Low Resolution Character Recognition with Increased Training Data and Image Enhancement

Shuhei Toba, Hirotaka Kudo, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

Proceedings of the 2014 Tohoku-Section Joint Convention of Institutes of Electrical and Information Engineers, Japan　2014/08/21

More details Close

2A18
Automatic Generation of Kanji Fonts from Sample Designs

Tatsunori Tsuchiya, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

Proceedings of the 2014 Tohoku-Section Joint Convention of Institutes of Electrical and Information Engineers, Japan　2014/08/21

More details Close

2A17
ぼけ除去及び複数フレームを利用した超低解像度文字認識

Shuhei TOBA, Hirotaka KUDO, Tomo MIYAZAKI, Yoshihiro SUGAYA, Shinichiro OMACHI

2014年画像の認識・理解シンポジウム (MIRU2014)　2014/07
サンプルデザインからの漢字フォントの自動生成

Tatsunori TSUCHIYA, Tomo MIYAZAKI, Yoshihiro SUGAYA, Shinichiro OMACHI

2014年画像の認識・理解シンポジウム (MIRU2014)　2014/07
効率的な映像符号化のための画像抽象化法の比較検討

Ryosuke ISHIMORI, Tomo MIYAZAKI, Yoshihiro Sugaya, Shinichiro Omachi

電子情報通信学会2014年総合大会学生ポスターセッション　251-251　2014/03
Character Region Extraction by Exploiting Auxiliary Lines

Shiori Kawamura, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

IEICE Technical Report　113　(431)　119-120　2014/02/06
Publisher: The Institute of Electronics, Information and Communication Engineers

More details Close

PRMU2013-152
Representative Graph Generation for Graph-Based Character Recognition Peer-reviewed

Tomo Miyazaki, Shinichiro Omachi

Journal of the Institute of Image Electronics Engineers of Japan　40　(3)　439-447　2011

DOI： 10.11371/iieej.40.439 　

ISSN： 1348-0316 0285-9831
Fast method for extracting representative graph from decorative character images Peer-reviewed

Tomo Miyazaki, Shinichiro Omachi

Proceedings - 2010 2nd IEEE International Conference on Network Infrastructure and Digital Content, IC-NIDC 2010　219-223　2010

DOI： 10.1109/ICNIDC.2010.5657776 　
Structure Extraction from Multiple Decorative Character Images by Common Supergraph

Tomo Miyazaki, Shinichiro Omachi, Hirotomo Aso

Proceedings of the Meeting on Image Recognition and Understanding 2009　506-512　2009/07/20

More details Close

IS1-16
Extraction of Representative Structure of Decorative Character Images Peer-reviewed

Tomo Miyazaki, Shinichiro Omachi, Hirotomo Aso

PROCEEDINGS OF THE 2009 CHINESE CONFERENCE ON PATTERN RECOGNITION AND THE FIRST CJK JOINT WORKSHOP ON PATTERN RECOGNITION, VOLS 1 AND 2　944-+　2009

DOI： 10.1109/CCPR.2009.5343952 　
“Silhouette Image Recognition with Weighted Minimum Common Supergraph” Peer-reviewed

Tomo Miyazaki, Shinichiro Omachi, Hirotomo Aso

Journal of the Institute of Image Electronics Engineers of Japan　38　(5)　640-647　2009

DOI： 10.11371/iieej.38.640 　

ISSN： 1348-0316 0285-9831
Strucure Extraction from Silhouette Images by Weighted Minimum Common Supergraph

Tomo Miyazaki, Shinichiro Omachi, Hirotomo Aso

Proceedings of the Meeting on Image Recognition and Understanding 2008　1408-1413　2008/07/29

More details Close

IS5-3
Extraction of Structure of Silhouette Images by Weighted Minimum Common Supergraph Peer-reviewed

Tomo Miyazaki, Shinichiro Omachi, Hirotomo Aso

Proceedings of the Second Korea-Japan Joint Workshop on Pattern Recognition　107　(281)　45-49　2007
Publisher: The Institute of Electronics, Information and Communication Engineers
ISSN： 0913-5685

More details Close

It is desired to recognize objects which are distorted or rotated in images. Since it is difficult to recognize those objects by template matching in images, recognition using graph has been studied. A graph extracted by medial axis transform to a digital silhouette image is not always a graph which represents the essential structure that distincts silhouettes in a category, because of noise and distortion. Our aim in this paper is to extract the essential structure of silhouette. We propose a method to extract the structure by weighted minimum common supergraph. To show the validity of the proposed method, experiments are carried out for categorizing silhouette images using the extracted structure.

Show all ︎Show first 5

Misc. 25

天吊り案内板を用いた屋内ナビゲーション手法の検討

竹村貴文, 菅谷至寛, 宮崎智, 大町真一郎

平成30年度電気関係学会東北支部連合大会講演論文集　2A07　2018
監視カメラを活用したユーザの実世界位置の推定

増保純平, 宮崎智, 菅谷至寛, 大町真一郎

平成30年度電気関係学会東北支部連合大会講演論文集　1I05　2018
超解像を用いた動画像符号化に関する検討

桑野拓朗, 宮崎智, 菅谷至寛, 大町真一郎

平成30年度電気関係学会東北支部連合大会講演論文集　1D14　2018
Multi-Frame Super Resolution Using 3D Convolution and RNN Prediction

黄希, 宮崎智, 菅谷至寛, 大町真一郎

2F12　2018
Accurate Region Extraction of Character in Scene Image with User Guided Auxiliary Lines

Tomo Miyazaki, Shiori Kawamura, Yoshihiro Sugaya, Shinichiro Omachi

画像ラボ　28　(10月)　37-42　2017/10
Publisher: JAPAN INDUSTRIAL PUBLISHING CO., LTD.
ISSN： 0915-6755
重要度を考慮した情景画像中における文字情報抽出

大平康太, 宮崎智, 菅谷至寛, 大町真一郎

電子情報通信学会総合大会学生ポスターセッション　ISS-SP-199　2017
スパースコーディングを用いた動画像符号化に関する検討

八重樫日菜子, 宮崎智, 菅谷至寛, 大町真一郎

電子情報通信学会総合大会学生ポスターセッション　ISS-SP-200　2017
Determining Important Objects in Scene Image Using Neural Networks

Tohoku-Section Joint Convention Record of Institutes of Electrical and Information Engineers, Japan　2017　2B12-158　2017
Publisher: Organizing Committee of Tohoku-Section Joint Convention of Institutes of Electrical and Information Engineers, Japan
DOI： 10.11528/tsjc.2017.0_158 　
Efficient Coding for Video Including Text Using Image Generation

Yosuke Nozue, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

57　(3)　2016/03/15

ISSN： 1882-7764
非線形最小化によるグラフのモデルの構築と画像認識

酒井利晃, 宮崎智, 菅谷至寛, 大町真一郎

電子情報通信学会2016年総合大会講演論文集　D-12-96　2016
Graph Learning with Quadratic Programming in Consideration of Class Diversity

Toshiaki Sakai, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

画像の認識・理解シンポジウム　PS2-42　2016
顕著性を利用した情景画像からの重要な文字列の検出

大平康太, 宮崎智, 菅谷至寛, 大町真一郎

画像の認識・理解シンポジウム　PS3-42　2016
パーツの生成による少数サンプルからのフォント生成

景山竣, 宮崎智, 菅谷至寛, 大町真一郎

画像の認識・理解シンポジウム　PS3-08　2016
Signboard Extraction and Recognition in Subway Station Premises

Tohoku-Section Joint Convention Record of Institutes of Electrical and Information Engineers, Japan　2016　1A03-3　2016
Publisher: Organizing Committee of Tohoku-Section Joint Convention of Institutes of Electrical and Information Engineers, Japan
DOI： 10.11528/tsjc.2016.0_3 　
Low Resolution Character Recognition Using Convolutional Neural Networks

Kyoko Maeda, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

平成28年度電気関係学会東北支部連合大会講演論文集　1A04　2016
文字認識を利用した環境中の日本語英字翻訳を行うウェアラブルシステムの開発

坂井清士郎, 宮崎智, 菅谷至寛, 大町真一郎

平成28年度電気関係学会東北支部連合大会講演論文集　2016　2G04-205　2016
Publisher: 電気関係学会東北支部連合大会実行委員会
DOI： 10.11528/tsjc.2016.0_205 　
Detection of a Key String from Scene Images Using Saliency

Kota Oodaira, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

International Workshop on Emerging ICT　2016
Investigation of Convolutional Neural Network Structure for Low Resolution Character Recognition

Kyoko Maeda, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

International Workshop on Emerging ICT　2016
Automatic Synthesis of Character Pattern Using Patch Transform

Jian Wang, Hiroya Saito, Tomo Miyazaki, Yoshihiro Sugaya, Shinichiro Omachi

Proceedings of the International Workshop on Electronics and Communications　2014
Rough Detection of Text in Scene Images by Finding Stroke Parts

Tomo Miyazaki, Shinichiro Omachi

Proceedings of the International Workshop on Electronics and Communications　2014
Extracting Representative Graph of Decorative Character Images by Random Method

Tomo Miyazaki, Shinichiro Omachi

Proceedings of the Third Student Organizing International Mini-Conference on Information Electronics Systems　67-68　2010
Iterative Extraction of Representative Graph Using Common Features from Decorative Character Images

Tomo Miyazaki, Shinichiro Omachi

Proceedings of the Second Student Organizing International Mini-Conference on Information Electronics Systems　85-86　2009
Extraction of Structure of Shapes Using Weighted Minimum Common Supergraph

Tomo Miyazaki, Shinichiro Omachi, Hirotomo Aso

Proceedings of the China-Korea-Japan Graduates Workshop on Electronic Information　49-50　2008
Structure Extraction from Silhouette Images by Weighted Minimum Common Supergraph

Tomo Miyazaki, Shinichiro Omachi, Hirotomo Aso

Proceedings of the First Student Organizing International Mini-Conference on Information Electronics Systems　75-76　2008
Silhouette Image Recognition

Tomo Miyazaki, Shinichiro Omachi, Hirotomo Aso

Proceedings of the Third Korea-Japan Joint Workshop on Pattern Recognition　13-14　2008

Show all ︎Show first 5

Presentations 1

Recent Advances in Image Processing by Artificial Intelligence International-presentation

Tomo Miyazaki

International Workshop on Emerging ICT　2018/11/05

More details Close

With the rise of deep learning, artificial Intelligent becomes a buzzword, which we see this word almost everyday in televisions, newspapers, blogs, or books. Artificial intelligence brought tremendous progresses to a wide of research areas. This talk describes what the artificial intelligence is and introduces recent advance in the field of image processing by the deep learning.

Industrial Property Rights 1

情報端末、位置推定方法、および位置推定プログラム

菅谷至寛, 外崎健人, 大町真一郎, 宮崎智

Property Type: Patent

Research Projects 10

人工知能を応用した自動麻酔制御システムの構築

星島宏, 宮崎智, 大町真一郎, 水田健太郎

Offer Organization: 日本学術振興会

System: 科学研究費助成事業基盤研究(C)

Category: 基盤研究(C)

Institution: 東北大学

2022/04/01 - 2025/03/31
Understanding of Essential Character Structure for Machine Learning and Kuzushiji Recognition

Offer Organization: Japan Society for the Promotion of Science

System: Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research (C)

Category: Grant-in-Aid for Scientific Research (C)

Institution: Tohoku University

2022/04/01 - 2025/03/31
Development of robotic sedation system

Offer Organization: Japan Society for the Promotion of Science

System: Grants-in-Aid for Scientific Research Grant-in-Aid for Challenging Research (Exploratory)

Category: Grant-in-Aid for Challenging Research (Exploratory)

Institution: Tohoku University

2021/07/09 - 2024/03/31
デフォルメされたマップでのユーザー指向な屋内ナビゲーションの実現

菅谷至寛, 宮崎智, 大町真一郎

Offer Organization: 日本学術振興会

System: 科学研究費助成事業基盤研究(C)

Category: 基盤研究(C)

Institution: 東北大学

2021/04 - 2024/03

More details Close

研究代表者らは、掲示されたフロアマップとスマートフォン内蔵のセンサだけで、施設管理者の協力がなくても屋内ナビを実現できる「ユーザー指向」な手法の研究開発を行ってきたが、デフォルメされ歪みの大きいマップも存在する。そのようなマップはこれまでの手法では対応できないことがわかっており、本研究課題ではユーザの負担にならない方法で歪みのあるマップに対応できる手法を開発することを目的としている。当該年度では、歪みのあるマップでの推定精度を改善するために、マップマッチングに用いる粒子フィルタの処理の改良を行った。交差点や曲がり角のマップ上での位置と、方向転換ステップの推定位置を利用して縮尺を求めて保存しておき、粒子の縮尺の候補として利用する。また、マップの縮尺が一様でない場合に、位置と縮尺の両方について同時に正しいパラメータを持つ粒子が存在する確率は非常に低いが、各々は正しい場合が少なくないという観察に基づき、方向転換時に粒子のパラメータをシャッフルする手法を提案した。これらにより、多くの歪みのあるマップでの位置推定精度が向上した。また、従来手法でもマップの小さなデフォルメに対応するために通路方向を利用していたが、広場領域において粒子の進みすぎ、進まなすぎが発生していた。この問題に対処するために、粒子の方位を補正する新しい手法を提案した。この手法では補正の閾値と量を通路の確信度によって動的に変化させ、無理のない補正を行う。これによって、従来よりも高い推定精度が得られることを実験によって確認した。
Development of Technology for Realizing High-Compression of Video by Extracting Important Regions

Offer Organization: Japan Society for the Promotion of Science

System: Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research (B)

Category: Grant-in-Aid for Scientific Research (B)

Institution: Tohoku University

2020/04 - 2023/03
Molecular Structure Analysis using Statistic Graph Model and Its Applications

Miyazaki Tomo

Offer Organization: Japan Society for the Promotion of Science

System: Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research (C)

Category: Grant-in-Aid for Scientific Research (C)

Institution: Tohoku University

2019/04 - 2022/03

More details Close

With the development of information and communications, various large-scale molecular science data have been constructed on the Internet. It is expected that useful knowledge (e.g., molecular structures that determine pharmacological activity) can be discovered from these large-scale molecular data. However, such discovery generally relies on experts, and the automatic extraction of essential structures from large-scale data is challenging. In this study, we developed a method to statistically and automatically analyze important structures that determine the chemical properties of molecules based on large-scale molecular data. As a major achievement, we developed a method to estimate the chemical properties of molecules with high accuracy by deep learning and showed that the method is effective for analyzing molecules in physical chemistry and other fields. Furthermore, the molecular analysis method was applied to image processing to estimate the fat content of fish.
カメラを持たないビジョン技術への挑戦

大町真一郎, 菅谷至寛, 宮崎智

Offer Organization: 日本学術振興会

System: 科学研究費助成事業挑戦的研究(萌芽)

Category: 挑戦的研究(萌芽)

Institution: 東北大学

2018/06 - 2021/03

More details Close

ユーザがカメラを持ち歩く代わりに監視カメラ等の既設の固定カメラを利用し、その映像とユーザの行動を結びつけることでビジョン技術を活用するための基盤技術を開発している。本年度は主に、固定カメラを利用してユーザの位置・姿勢を推定する手法について検討した。具体的には、実世界に対応した俯瞰マップを用意し、動画像上の座標をマップ座標に変換をすることで位置推定を行う。動画像中の人物位置を、物体検出手法を応用することで検出する。物体検出には、ニューラルネットワークを用いた手法を採用し、学習によって精度を向上させる。そして、その人物の足元の座標を座標変換により求め、立っている位置を特定する。さらに、人物の視線を知るために、頭部方向を推定する。そのためにウェアラブルセンサの出力を用いる。メガネ型デバイスを用い、このデバイスに内蔵されているセンサの出力により人物の異動の加速度および角速度を推定し、頭部の方向を推定する。そして、実際に研究室内の環境で測定を行い、位置推定および頭部方向推定がある程度可能であることを確認した。また、これとは別に、環境中の文字認識を利用した情報提供アプリケーションのためのウェアラブルシステムについて検討した。認識結果の提示方法、ユーザインタフェース、どのようなアプリケーションがユーザにとって有益かなどについて、主観評価実験を行い評価した。その結果、ユーザに提示する情報およびユーザインタフェースの設計についてある程度の見通しを得た。
Pedestrian navigation everywhere: Development of pedestrian navigation without prior data collection

SUGAYA Yoshihiro

Offer Organization: Japan Society for the Promotion of Science

System: Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research (C)

Category: Grant-in-Aid for Scientific Research (C)

Institution: Tohoku University

2018/04 - 2021/03

More details Close

In this research, we aimed to establish a technique for realizing pedestrian navigation using an information board, which is provided for humans. Although conventional methods had significant limitations on the floor maps that can be used, we introduced deep learning to deal with various types of maps. We also proposed a method to determine the initial parameters using known objects in the map, and showed the possibility of eliminating the manual specification of the initial position by the user. In addition, we studied the position estimation using a ceiling-suspended guide plate.
Development of Technology for Realizing High-Compression of Video while Ensuring Visibility

OMACHI Shinichiro

Offer Organization: Japan Society for the Promotion of Science

System: Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research (B)

Category: Grant-in-Aid for Scientific Research (B)

Institution: Tohoku University

2016/04 - 2019/03

More details Close

We regard the quality of a video as the visibility of important information, and developed a technology to realize high compression rate without losing the value as video and necessary information. There are two major contributions. One is the detection method of the important region. We proposed a method using visual saliency and a method using the analysis of the meaning. The other is an efficient compression method. We proposed a method using sparse coding, a method applying image generation technology, and a method using super resolution.
Flexible and Accurate Recognition for Non-Rigid Object using Graph Matching

Miyazaki Tomo

Offer Organization: Japan Society for the Promotion of Science

System: Grants-in-Aid for Scientific Research Grant-in-Aid for Research Activity start-up

Category: Grant-in-Aid for Research Activity start-up

Institution: Tohoku University

2015/08 - 2017/03

More details Close

Graphs describe non-rigid objects which vary greatly and flexibly. However, graphs are not used for pattern recognition for image objects due to the following two problems: difficulty in extracting graphs from images and lack of a method for measuring similarity of graphs. In this study, we propose a method for image object recognition by applying probabilistic graph model to measure similarity of graphs extracted from feature points in an image. In addition, we show the improvement of recognition performance using several probabilistic graph models. These results are significant in not only patter recognition society but also industry because a use of graphs can be facilitated by the proposed method.

Show all Show first 5

Teaching Experience 5

電気回路学基礎論
電気回路学基礎演習
アドバンス創造工学研修
創造工学研修
Database　Tohoku Institute of Technology

Social Activities 2

秋田高校出張講義

2024/05/29 - 2024/05/29
人工知能と画像処理技術

高大連携事業に係る地域公開講座

2023/10/17 - 2023/10/17

Academic Activities 6

Publicity Chairs, The International Conference on Document Analysis and Recognition (ICDAR)

2017 -

Activity type: Competition, symposium, etc.
Program Committee Member, International Workshop on Historical Document Imaging and Processing (HIP)

2015 - Present

Activity type: Competition, symposium, etc.
Program Committee Member, IJCAI-PRICAI

2020 - 2021

Activity type: Academic society, research group, etc.
組織副委員長第23回画像の認識・理解シンポジウム

2020 -
MIRU2019 若手プログラム委員

2019 -

Activity type: Competition, symposium, etc.
Program Committee Member, The 16th International Conference on Frontiers in Handwriting Recognition (ICFHR)

2018 -

Activity type: Competition, symposium, etc.

Show all Show first 5